Optimal caching in content-centric mobile networks using file segmentation

Abstract

In this paper, we study the throughput performance in a content-centric mobile ad hoc network, where each node moves using multihop according to the random walk mobility model and requests a file from the library independently at random, according to a Zipf popularity distribution. In our network model, we assume that each file consists of K distinct segments of equal size such that each of  mobile nodes is able to transmit completely one segment to one of its neighbor cells in one time slot. Using multihop transmission, the throughput scaling law is analyzed based on the following two reception strategies, which are introduced to determine how K distinct segments are fully delivered to the requesting node in turn to rebuild the desired file: A sequential reception and a random reception. Then, we analyze the throughput of the content-centric wireless networks and find the optimal cache allocation strategies analytically, which maximize the throughput. Our main result indicates that the throughput obtained from the random reception strategy outperforms the sequential reception case.